W3C home > Mailing lists > Public > www-rdf-comments@w3.org > October to December 2003

Re: entailment-from-inconsistent-graph [was: proposed test of RDFS entailment rules]

From: pat hayes <phayes@ihmc.us>
Date: Sat, 18 Oct 2003 09:18:00 -0400
Message-Id: <p06001f12bbb6e6b0c434@[192.168.0.38]>
To: "Jos De_Roo" <jos.deroo@agfa.com>
Cc: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>, www-rdf-comments@w3.org, www-rdf-comments-request@w3.org

The current (editor's draft) RDFS rules have a 
criterion for detecting inconsistency, to wit, 
the derivation of a triple called an 'XML clash'.

_:nnn rdf:type rdfs:Literal .

where the subject bnode _:nnn was introduced, and 
allocated to an ill-typed literal by, the lg 
generalization rule (formerly called rdf2). The 
derivation for this example is as follows:

<http://example.org/prop> 
<http://www.w3.org/2000/01/rdf-schema#range> 
<http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral> 
.

<http://example.org/foo> 
<http://example.org/prop> 
"<"^^<http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral> 
.

<http://example.org/foo> 
<http://example.org/prop> _:1*.      rule lg, 
with _:1* allocated to 
"<"^^<http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral>, 
which is ill-typed. (Jos, can your code keep 
track of this when the rule is applied and 'mark' 
the bnode accordingly?)

_:1*
<http://www.w3.org/1999/02/22-rdf-syntax-ns#type>
<http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral> .   rule rdfs3

<http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral>
<http://www.w3.org/2000/01/rdf-schema#subClassOf>
<http://www.w3.org/2000/01/rdf-schema#Literal> . 
RDFS axiomatic triple

_:1* <http://www.w3.org/1999/02/22-rdf-syntax-ns#type>
<http://www.w3.org/2000/01/rdf-schema#Literal>  .               rule rdfs9

So the 'ex cont quod' could reasonably be 
restricted to this case, ie if you really believe 
an XML clash then you will believe anything. 
However, notice that the clash itself is not 
inconsistent: it is a symptom of the original set 
being inconsistent. So it would not be correct to 
say that the silly conclusion is entailed by the 
clash; rather, if you can derive a clash from a 
graph, then  the silly conclusion is entailed by 
your original graph.

>What a coincidence - while sitting in a plane this evening
>I did't think to implement a "ex contradictione quodlibet".
>The premise graphs are assumed to be the case unless they
>can be proven to be inconsistent and then we just say so
>and don't explicitly use them further. So we can't
>run that testcase.

I think it would be OK to be able to prove the 
antecedent inconsistent, and call that a proper 
run of the test-case. I think that was Peter's 
main point.

Pat

>
>
>--
>Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/
>
>
> 
>                                                                                                                                        
>                       Brian 
>McBride                                                                                                      
>                       <bwm@hplb.hpl.hp.com 
>To:       "Peter F. Patel-Schneider" 
><pfps@research.bell-labs.com>                     
>                       > 
>cc: 
>www-rdf-comments@w3.org                                                      
>                       Sent by: 
>Subject:  entailment-from-inconsistent-graph 
>[was: proposed test of RDFS entailment    
>                       www-rdf-comments-req 
>rules]                                                                                
> 
>uest@w3.org                                                                                                        
> 
>                                                                                                                                        
> 
>                                                                                                                                        
>                       2003-10-15 03:37 
>PM                                                                                                
> 
>                                                                                                                                        
> 
>                                                                                                                                        
>
>
>
>
>
>Peter,
>
>The WG were unable to discuss this suggestion before publishing the 2nd
>last call documents.  I propose to track this as a 2nd last call comment:
>
>http://www.w3.org/2001/sw/RDFCore/20031010-comments/#entailment-from-inconsistent-graph
>
>
>Brian
>
>
>
>
>Peter F. Patel-Schneider wrote:
>>  I propose that the following be a positive entailment test in the RDF
>test
>>  suite.  This is a valid RDFS entailment (modulo typing errors), but is
>not
>>  a consequence of the current RDFS entailment rules.
>>
>>  Premise
>>
>  > <http://example.org/prop> <http://www.w3.org/2000/01/rdf-schema#range> <
>http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral> .
>>  <http://example.org/foo> <http://example.org/prop> "<"^^<
>http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral> .
>  >
>>  Conclusion
>>
>>  <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <
>http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <
>http://www.w3.org/1999/02/22-rdf-syntax-ns#type> .
>>
>>
>>  Peter F. Patel-Schneider


-- 
---------------------------------------------------------------------
IHMC	(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32501			(850)291 0667    cell
phayes@ihmc.us       http://www.ihmc.us/users/phayes
Received on Saturday, 18 October 2003 09:16:35 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 21 September 2012 14:16:33 GMT